I was introduced to Erlang at Rdio (this shaped the future of my career) and worked on the original implementation of the Heavy Rotation feature. I walked around the office asking people "Does this look right? How about this?" I'd upgrade Erlang code in production by hot-reloading the heavy rotation module, and walk around some more. It was fun!
In 2011, I realized that Rdio will never succeed due to lack of fanatical leadership—you can't have a company with absent founders, that's just not how it works—and left. I've been expecting this sort of an "exit" for a while.
Today I work at my own startup (sameroom.io) and every single day I'm grateful for the Rdio experience—mostly as a cautionary tale, unfortunately.
And now, we have to deal with this:
"Are we really going to start using Spotify after so many years of hating on it?" —our 15-year-old.